New research on philoxeroides (alligator weed) in its South American native range

Alejandro J. Sosa,1 Mic H. Julien2 and Hugo A. Cordo1

Summary Alternanthera philoxeroides (alligator weed) is a herbaceous amphibious weed of the Amaranthaceae, native to southern South America. Several agents from , e.g. hygrophila and , have been used to control aquatic A. philoxeroides in Australia and the USA. However, in Australia, the weed continues to pose a serious problem, particularly in terrestrial situations. In Argentina, A. philoxeroides is distributed along the catchments of the Paraná and Uruguay rivers in the north, and in the catchments of the San Borombón and Salado rivers in the centre of prov- ince. Two forms are recognized: A. philoxeroides f. philoxeroides in the southern range and A. philoxeroides f. angustifolia in the northern range. There appears to be preferential attack by flea on A. philoxeroides f. angustifolia. In 2000, the CSIRO initiated a collaborative research project with the USDA South American Biological Control Laboratory in Argentina to search for new biolog- ical agents. After the initial year of surveys, the natural enemies that may have biological control poten- tial included: two species of leaf-feeding beetles, Systena spp.; a tip-galling Cecidomyiidae fly; and two agromyzid flies, one that causes node galls and another that mines leaves. Two fungi were also found: one probably Nimbya alternantherae, known to have a wide host range, and another, thought to be a new Sphaceloma species, that causes a characteristic “corky” deformation on the stem and leaf surfaces. Surveys will be extended and the interactions between these herbivores and pathogens with A. philoxeroides will be studied.

Keywords: alligator weed, Alternanthera philoxeroides, native range, natural enemies.

Introduction moth were subsequently released in Australia and controlled the weed in warm temperate aquatic habitats In its introduced range, Alternanthera philoxeroides (Julien 1981). However, terrestrial growth of A. (Martius) Grisebach (alligator weed; Amaranthaceae) philoxeroides and aquatic growth in cooler regions of is often a serious aquatic and terrestrial weed. Surveys Australia continue to cause serious concern (Julien & for natural enemies were carried out in the 1960s in Bourne 1988, Julien & Stanley 1999). parts of its native range in South America (Vogt 1961), Alternanthera philoxeroides consists of several taxa and the flea Selman and in both its native and adventive ranges. It was first Vogt, the moth Arcola malloi (Pastrana) and the described by Martius in 1826, and named Bucholzia Amyinothrips andersoni O’Neil were released in the philoxeroides. Covas (1939, 1941) considered two vari- USA. Good control was obtained in aquatic habitats, eties of alligator weed: A. philoxeroides var. obstusifolia largely attributed to the and the moth (Moquin) Hicken and A. philoxeroides var. acutifolia (Spencer & Coulson 1976). The flea beetle and the (Moq.) Hicken. The former was characterized by the presence of ovate lanceolate leaves with obtuse or sub obtuse apex; whereas the latter variety, acutifolia, has 1 South American Biological Control Laboratory, USDA-ARS, Bolivar lanceolate leaves with acute apex. Pedersen (1967) indi- 1559 (B1686EFA), Hurlingham, Buenos Aires, Argentina. cated that Martius did not consider two varieties of alli- 2 CSIRO Entomology, Long Pocket Laboratories, 120 Meiers Road, Indooroopilly, Queensland 4068, Australia. gator weed and he considered that the specimen Corresponding author: A.J. Sosa . deposited in Brussels, which had ovate or elliptic and

180 New research on alligator weed in South America obtuse leaves, as a good lectotype. However, Pedersen located from Posadas, on the northern Argentina (1999) suggested that A. philoxeroides be divided into with , to , well south of Buenos Aires. two and he referred to these as forms: A. philoxeroides f. In the field, the presence of fruits and seedlings of A. philoxeroides (Mart.) Griseb. and A. philoxeroides f. philoxeroides was recorded, and collected for cultiva- angustifolia Süssenguth. tion in the laboratory. Seedlings and stem cuttings were Records from the United States suggest that both collected from four localities and also grown in the forms were introduced there. Ganstad & Solymosy laboratory. Two localities represented the southern (1973) and Weldon et al. (1973) reported that some form, Tandil and Mar del Plata in Buenos Aires prov- plants of A. philoxeroides had small stems, and in some ince, and two represented the northern form, Santa Fé, cases the internodes were solid, and apparently this Santa Fe , and Hurlingham, Buenos Aires tended to produce a deficient plant that is not preferred province. After 6 months of growth in identical condi- by the flea beetle A. hygrophila. Kay & Haller (1982) tions, 27 plant parameters were measured and found two different forms in the USA, which they compared using principal components analysis. Three differentiated as the narrow-stemmed alligator weed factors were extracted that explained 71% of the total (NSA) biotype and the broader-stemmed alligator weed variance. They were: diameter of internode of leaf one, (BSA) biotype. The NSA biotype was characterized by length/width ratio of leaf one, and leaf apical angle of the presence of slender stems with short internodes and leaf one. ANOVA was carried out to evaluate mean obtuse and rounded leaves. The BSA biotype has differences and Tukey test was used for multiple broader and longer stems and longer and acute leaves. comparisons among pairs of means based on unequal They pointed out a different pattern in the damage sample of sizes. All statistical analysis was carried out caused by A. hygrophila. Populations of NSA seem to using Statistica 5.5. be attacked less than the BSA populations in the USA and both biotypes responded differently to herbicides (Kay 1992). Wain et al. (1984) demonstrated genetic Results differences between the two biotypes in the USA using isozyme pattern analyses. In contrast, no genetic varia- Alligator weed: native ranges, tion was detected within or between populations of A. morphological variation and biology philoxeroides in China using RAPD analysis (Xu et al. Alternanthera philoxeroides f. philoxeroides, the 2003). southern form, was distributed along the catchments of In 2001, the CSIRO (Australia) initiated a coopera- San Borombón and Salado rivers in Buenos Aires prov- tive research project with the USDA South American ince (Fig. 1). It was not found in western Buenos Aires Biological Control Laboratory in Argentina to update province. It was also found in some sites in the north the list of natural enemies known from A. philoxeroides west of Argentina, in township drains, possible outside and to identify potential new biological control agents. the native range. This form has small, ovate leaves, short internodes and slender stems. It occurred in semi- Materials and methods aquatic conditions, in ditches next to roads or along the shores of lakes. Flowering was abundant in summer, Surveys were conducted, mostly in Argentina, but also with short, small inflorescences. in Uruguay, Paraguay and south-eastern Brazil, The second, northern form, A. philoxeroides f. between October 2001 and November 2002. Natural angustifolia, was distributed along the Paraná River enemies were sampled at 93 A. philoxeroides sites. from Posadas and along the from Sites where Alternanthera aquatica (Parodi) Chodat Pantanal region (Brazil). It also occurred along the (ex A. hassleriana) occurred were also sampled. At Uruguay River downstream from Santo Tomé (Corri- each A. philoxeroides site, adult were collected entes province). Thus, it extended from the northern either by direct aspiration from plants or after sweeping parts of through the north-east with a net, placed in 70% ethanol and sent to taxono- wetlands of Argentina (including Entre Rios and Corri- mists for identification. Other immature insects were entes province, western Misiones and eastern parts of collected alive and reared in the laboratory to adult the that have their eastern along the stage. These were also sent for identification. Paraná and Paraguay rivers), and along the south and The extent of the native ranges of both forms of A. east coast of Uruguay and possibly into south-east philoxeroides were approximated using the local distribu- coastal areas of Brazil. This form was also found in the tion of the weed, its morphology, and assessments of north-west of Argentina and at one site in the north of where it grew, i.e. natural areas or highly disturbed loca- Patagonia in Río Negro province (Fig. 1). In these cases tions such as town drains. Herbarium specimens from they were only associated with human activities and representative sites, six stems per site, were collected. For were considered to be outside the native range. Vogt et each of the six stems, two younger leaves were removed al. (1979) suggested that north-western populations and stored in 96% alcohol for genetic analysis. RAPD were isolated relicts or introduced populations. We analyses were conducted on material from seven sites suggest the latter.

181 Proceedings of the XI International Symposium on Biological Control of Weeds

70° 65° 60° 55° 50° 45° 40° 15° 15° Native range of alligator weed A. philoxeroides f. philoxeroides A. philoxeroides f. angustifolia A. philoxeroides “Santa Fé” 20° 20°

Paraguay Brazil Chile

25° 25°

Argentina 30° 30°

Uruguay Atlantic Ocean

35° 35°

km

0 200 400 40° 40° 70° 65° 60° 55° 50° 45° 40°

Figure 1. Distribution of Alternanthera philoxeroides in Argentina, Paraguay, Uruguay and the south-eastern coast of Brazil. Closed lines indicate the native range for A. philoxeroides f. philoxeroides in the south and broken lines indicate the native range of A. philoxeroides f. angustifolia in the north and east.

The growth comparison experiments distinguished Sexual reproduction in alligator weed the two forms, but also indicated the possibility of a third Seeds and, for the first time, seedlings were observed form. Alternanthera philoxeroides f. philoxeroides, in the field. Initially these were found at several sites grown from material from Tandil and Mar del Plata, had near the at Dolores and along Route 30 significantly more slender stems (smaller diameter inter- between Azul and Tandil, Buenos Aires province, from nodes) and ovate (higher values in length/width ratio and the philoxeroides form. Later, seeds were collected from leaf apical angle) and smaller leaves compared with the the angustifolia form from Chaco province and were northern form, A. philoxeroides f. angustifolia. The latter germinated in the laboratory. Laboratory garden plants grown from Hurlingham material, was significantly of both forms also set viable seeds. Vogt (1961) differentiated by its acute and long leaves (less value in collected seeds from San Miguel del Monte, Buenos leaf apical angle and length/ width ratio) and broader Aires province, probably from the philoxeroides form, stems. This form was more frequently attacked by the and germinated these in small tins while travelling. flea beetle A. hygrophila. Plants grown from Santa Fé material resembled the angustifolia form by having broader stems; and the Natural enemies – insects philoxeroides form by having ovate and smaller leaves. A range of natural enemies was found (Table 1) and The Santa Fé site is located just outside the overlap some could be considered promising because of their regions of the sympatric distributions for the two forms abundance, damage to the plant and their probable (Fig. 1). It is therefore uncertain if plants from this narrow host range. The chrysomelid beetles, locality represent a third form or a hybrid. The RAPD Agasicles spp., that feed on leaves and stems, are analyses of samples from seven sites that included among those promising candidates. They were material representing both recognized forms of A. collected at 40 sites, mostly on A. philoxeroides f. philoxeroides gave ambiguous results. Further analyses angustifolia, and most specimens were identified as using AFPL technique are planned. A. hygrophila, the known successful biological control

182 New research on alligator weed in South America agent. Some of the specimens were identified as A. out that one of them (found in Posadas, Misiones prov- vittata Jacoby and A. conexa (Boheman) based on Vogt ince and in Barranqueras, Chaco province, Argentina), et al. (1979) and Selman & Vogt (1971). The three was the most abundant. These areas were searched in this species of Agasicles are very similar, difficult to distin- study and specimens from this genus were found at both. guish and inconsistencies are apparent. A revision of In Barranqueras and near Corumbá (Pantanal, Brazil) this genus would assist in determining the number of another species of Systena was found on A. aquatica. species involved and how to differentiate them. The thrips, andersoni O’Neill was found Disonycha argentinensis Jacoby was found at 24 at every site at every visit. This is the most ubiquitous sites, mostly in conjunction with Agasicles, except in San on A. philoxeroides regardless of plant form. It has Miguel del Monte (Buenos Aires province) where this also been observed attacking A. aquatica. The abun- flea beetle seemed to be dominant. This beetle was dance of this thrips, its presence throughout the plant’s studied and released as a possible biological agent for the growth cycle, in terrestrial and aquatic habitats, and its terrestrial A. philoxeroides in Australia and New Zealand known host specificity strongly suggest it as a promising but failed to establish, apparently due to environmental candidate for Australia. It has been introduced to the reasons (Julien & Chan 1992, Julien & Griffiths 1999). USA where it became established, but apparently has not Large populations of one species of Systena were contributed to the control of A. philoxeroides (Julien & found at 20 sites, mainly in Santa Fé and Chaco prov- Griffiths 1999). Details of its biology and rearing inces and its biology is being studied in the laboratory. methods are recorded in Vogt 1961, Maddox et al. 1971, Vogt (1961) reported three species of Systena, consid- Maddox & Mayfield 1972 and Maddox 1973. ered them as “minor biotic suppressants”, and pointed

Table 1. Natural enemies of Alternanthera philoxeroides found in Argentina. Species Form of A. Observations philoxeroides Chrysomelidae: Alticinae Agasicles hygrophilaa,b angustifolia Specific Agasicles conexa angustifolia A revision of this genus is Agasicles vittata angustifolia necessary Systena spp. angustifolia Being studied Disonycha argentinensis angustifolia Specific Paranapiacaba significata angustifolia Polyphagous species, philoxeroides considered a pest Thysanoptera: Amynothrips andersonib angustifolia Specific philoxeroides : Phyticidae Arcola malloia,b angustifolia Specific philoxeroides “leaf-tying moth” angustifolia Host-range unknown philoxeroides Diptera: Cecidomyiidae Clinodiplosis alternantherae angustifolia Probably specific philoxeroides Being studied Agromyzidae Ophiomyia alternantherae angustifolia Probably specific philoxeroides Being studied Ophiomyia marellii angustifolia Biology unknown Ophiomyia possibly buscki philoxeroides Host-range unknown Cicadellidae: Typhlocibinae Empoasca curveola angustifolia Not specific Empoasca aculeata philoxeroides Not specific Membracidae Membracidae sp. 1 angustifolia Host range unknown Fungi Nymbia alternantherae ? angustifolia Not specific philoxeroides Sphaceloma ? angustifolia Host range unknown philoxeroides a Species already released and established in Australia and USA. b Species already released and established in the USA.

183 Proceedings of the XI International Symposium on Biological Control of Weeds

The fly Clinodiplosis alternantherae n. sp. Gagné trial and aquatic habitats. When growing in terrestrial was abundant in Buenos Aires province and at several conditions they are morphologically similar and it is sites in Uruguay, and was also present through most of difficult to differentiate them (Bona & Lange de the native range of A. philoxeroides. The mesophyll and Morretes 1997). Some insects attack both species; for main vein of leaves are enlarged due to larval activity. example, the tip galler C. alternantherae and a flea In many cases the gall causes severe stunting of the beetle Systena sp. were observed on both species in inflorescence peduncle. We do not know if the gall Pantanal, Brazil, and Chaco province, Argentina. interferes with the production of seeds or with the size Alternanthera philoxeroides exhibits morphological and numbers of flowers. The fly is abundant, has a short variation. The differences observed between A. life cycle, is multivoltine and appears to be restricted to philoxeroides growing in aquatic or terrestrial habitats A. philoxeroides and the closely related A. aquatica. may be due to phenotypic expressions. However, the Vogt (1961) reported a species that formed terminal existence of at least two genetic entities under the name galls, but provided no other information. Gagné (1994) A. philoxeroides appears to be possible, but requires reported a species of Clinodiplosis that damages the confirmation. This, along with the existence of geneti- terminals of A. philoxeroides in Argentina, Brazil and cally different form(s) in Australia and other countries, Uruguay. Specimens have been recently sent to Gagné is currently being assessed. Such information may be and a paper with descriptions of this fly, as C. alternan- important for management of the weed, as genetically therae, is in preparation. different forms may respond differently to management Three species of Ophiomyia were found and prob- strategies. Preferential attack by A. hygrophila on ably correspond to those agromyzids mentioned by different morphotypes of A. philoxeroides in Argentina Vogt (1961). They are: O. alternantherae (Spencer), O. and the USA is already known. marellii (Brethes) and Ophiomyia possibly buscki The two currently recognized forms of A. (Frost). Ophiomyia alternantherae, the smallest one, is philoxeroides in Argentina produce fertile seeds. In the a leaf miner that was found in many (37) of the locali- USA, Ganstad & Solymosy (1973) collected seeds, but ties visited, and was frequently parasitized by wasps. they failed to germinate unless the utricle was There is very little information about its biology and the extracted. They were not able to identify pollination damage it causes to plants. Its probable short life cycle processes. Alternanthera philoxeroides has not been and its abundance in the field make it a potential candi- observed producing seeds in other exotic ranges (Julien date for biological control of A. philoxeroides. Ophio- 1995). The factors, if any, preventing development of myia marellii forms node galls and appears to be seeds in the exotic range must be studied. Alternanthera restricted to A. philoxeroides. It probably has a long life philoxeroides grows along much of the east coast of cycle. Its galls were mostly found on underground Brazil more or less contiguously with the Argentina/ stems but on a few occasions they were found just Uruguay populations. However, it is generally consid- above ground. This species was found at only seven ered that the native range centres around river systems sites, mostly on the angustifolia form and exclusively in Argentina (Vogt et al. 1979), and that the Brazil in terrestrial situations. The larvae of Ophiomyia populations are part of the extended range. possibly buscki are stem-miners. A few specimens only Several host-specific agents are known from A. were collected in one location near Tandil, Buenos philoxeroides: A. andersoni, D. argentinensis and Aires province. A taxonomic revision of this species is Agasicles spp. The biology and host ranges of other necessary because discrepancies appeared between potential agents are being studied in Argentina. They specimens collected and those deposited in museums. include C. alternantherae, Systena spp. and Ophiomyia spp. There is no doubt that A. andersoni is host specific Natural enemies – fungi on A. philoxeroides, causing intensive damage. The The pathogen, Nymbia alternantherae (Hyphomyc- lack of contribution by this insect to the control of A. etes) was found at most locations (82). This species has philoxeroides in the USA should not discourage its been studied in Brazil and the USA and has potential as release in Australia. Many agents work well in one a mycoherbicide. It has recently been found in Australia habitat, region or country, but not in another. (B. Auld, R. Gilbert & B. Hennecke, pers. comm.) Disonycha argentinensis has been studied previously where further studies will be conducted. A species of and its host range and rearing techniques are known. Sphaceloma (Coelomycete) was observed at 27 sites on This insect was found at most sites in the northern both forms of A. philoxeroides. Damaged parts of stems distribution of A. philoxeroides. It is considered a and leaves take on a cork-like texture and greyish potential candidate for the control of terrestrial A. colouration. Very little is known about this pathogen. philoxeroides and a renewed effort to import and estab- lish it in Australia is suggested. Discussion Once the Agasicles is clarified, biology and preliminary host-range studies can be carried out Two species of Alternanthera were found in Argentina: for species other than A. hygrophila, which has already A. philoxeroides and A. aquatica. Both grow in terres- been studied. Additionally, more information is

184 New research on alligator weed in South America required about the pathogens of A. philoxeroides, Julien, M.H. & Chan R.R. (1992) Biological control of alligator including more detailed field surveys to be conducted weed: unsuccessful attempts to control terrestrial growth by a plant pathologist, and biology and host-range using the flea beetle Disonycha argentinensis (Col.: Chrys- studies on the Sphaceloma sp. Native range studies will omelidae). Entomophaga 37, 215–221. be continued and surveys for new natural enemies will Julien, M.H. & Griffiths, M.W. (1999) Biological Control of Weeds. A World Catalogue of Agents and their Target be extended. Weeds, Fourth Edition. CABI International, Wallingford. Julien, M.H. & Stanley, J.N. (1999) The management of alli- Acknowledgements gator weed, a challenge for the new millennium. Proceed- ings of the 10th Biennial Noxious Weeds Conference (eds R. We thank J. Dorado and M.C. Hernandez for their Ensbey, P. Blackmore & A. Simpson), pp 2–13. NSW Agri- invaluable assistance. We appreciate the help of N. culture, Australia. Cabrera and S. Paradell from Museo de Ciencias Natu- Maddox, D.M. (1973) Amynothrips andersoni (Thysanoptera: rales de La Plata (Argentina) for identifying Chrys- Phlaeothripidae), a thrips for the biological control of alli- omelidae and Cicadellidae, respectively, R. Gagné gator weed. 1. Host specificity. Environmental Entomology from the Systematic Entomology Laboratory (USA) for 2, 30–37. identification of the Cecidomyiidae, G. Valladares Maddox, D.M., Hinnesey, L.A., Blackburn, R.D, & Spencer, from Univesidad Nacional de Cordoba (Argentina) for R.D. (1971) Insects to control alligator weed, an invader of aquatic ecosystems in the United States. BioScience 21, identification of Agromyziidae, and R. Barreto from 985–991. Universidade Federal de Viçosa (Brazil) for identifica- Maddox, D.M. & Mayfield, A. (1972) A method for rearing and tion of pathogens. This work was funded by the Natural studying Amynothrips andersoni in the laboratory. Journal Heritage Trust (Australia). of Economic Entomology 65, 1521–1523. Pedersen, T.M. (1967) Studies in South American Amaran- References thacea. Darwiniana 14, 430–463. Pedersen, T.M. (1999) Amaranthaceae. Catálogo de las Plantas Bona, C. & Lange de Morretes, B. (1997) Anatomia compara- Vasculares de la Argentina II (eds F.O. Zuloaga & O. tiva do caudale de Alternanthera philoxeroides (Mart.) Morrone), pp 12–31. Monographs in Systematic Botany Griseb. e A. aquatica (Parodi) Chodat. (Amaranthaceae). from the Missouri Botanical Garden. Arquivos de Biologia e tecnología 40, 285–296. Selman, B.J. & Vogt, G.B. (1971) Lectotype designations in Covas, G. (1939) Los géneros de amarantáceas . South American genus Agasicles (Coleoptera: Chrysomel- Revista Argentina de Agronomía 6, 282–239. idae), with descriptions of a new species important as Covas, G. (1941) Las Amarántaceas Bonarienses. Darwiniana suppressant of alligator weed. Annals of the Entomological 5, 329–368. Society of America 64, 1016–1020. Gagné, R.J. (1994) The gall midges of the Neotropical Region. Spencer, N.R. & Coulson, J. (1976) The biological control of Cornell University Press. New York. Alternanthera philoxeroides, in the United States. Aquatic Ganstad, E.O. & Solymosy, S.L. (1973) Control of alligator Botany 2, 177–1990. weed in Louisiana and the gulf coast area. Biological Vogt, G.B. (1961) Exploration for natural enemies for alligator Control of Alligator Weed. United States Army Engineer weed and related plants in South America. Unpublished Waterways Experiment Station, Technical Report 3, pp internal report, USDA, special report PI–5. A1–A23. United States Army Engineer Waterways Experi- ment Station, Vicksburg, Mississippi. Vogt, G.B., McGuire, J.U., Jr. & Cushman, A.D. (1979) Prob- Kay, S.H. (1992) Response of two alligator weed biotypes to able Evolution and Morphological Variation in South Amer- quinclorac. Journal of Aquatic Plant Management 30, ican Disonychine Flea Beetles (Coleoptera: 35–40. Chrysomelidae) and Their Amaranthaceous Hosts. USDA Kay, S.H. & Haller, W.T. (1982) Evidence for the existence of Technical Bulletin No. 1593. distinct alligator weed biotypes. Journal of Aquatic Plant Wain, R.P., Haller, W.T. & Martin, D.F. (1984) Genetic rela- Management 20, 37–41. tionship among two forms of alligator weed. Journal of Julien, M.H. (1981) Control of aquatic Alternanthera Aquatic Plant Management 22, 104–105. philoxeroides in Australia; another success for Agasicles Weldon, L.W., Blackburn, R.D. & Durden, W.C., Jr. (1973) hygrophila. Proceedings of the V International Symposium Evaluation of the Agasicles n. sp. for biological control of on Biological Control of Weeds (ed E.S. Delfosse), pp alligator weed. Biological Control of Alligator Weed. United 583–588. CSIRO, Melbourne. States Army Engineer Waterways Experiment Station, Tech- Julien, M.H. (1995) Alternanthera philoxeroides (Mart.) nical Report 3, pp. D1–D54. United States Army Engineer Griseb. The Biology of Australian Weeds (eds R.H. Groves, Waterways Experiment Station, Vicksburg, Mississippi. R.C.H. Shepherd & R.G. Richardson), Volume 1, pp 1–12. Xu, C-Y., Zhang, W-J., Fu, C-Z., & Lu, B-R. (2003) Genetic Julien, M.H. & Bourne, A.S. (1988) Alligator weed is spreading diversity of alligator weed in China by RAPD analysis. in Australia. Plant Protection Quarterly 3, 91–96. Biodiversity and Conservation 12, 637–645.

185